Society: AGA

Background: Recurrent Clostridioides difficile infection (rCDI) is a debilitating disease associated with significant healthcare burden and low quality of life. The greatest risk factor for rCDI is a history of recurrence, although older age (≥65 years) and comorbidities also increase risk. Recurrent CDI occurs in 20-36% of patients with 1st recurrence and ≥40% in those with ≥2 recurrences. In a Phase 3 randomized controlled trial (ECOSPOR III), the investigational oral microbiome therapeutic SER-109 was superior to placebo in reducing risk of rCDI at 8 weeks in patients with ≥2 recurrences (12% vs 40%) and this response was durable through 24 weeks [Feuerstadt NEJM 2022; Cohen JAMA 2022]. Here we present safety and efficacy data from a Phase 3 open-label trial (ECOSPOR IV), including durability of response in patients with 1^st and ≥2 recurrences.

Methods: Adults with rCDI were enrolled at 72 North American sites in 2 cohorts: 1) rollover subjects with rCDI in ECOSPOR III, diagnosed by toxin EIA) and 2) subjects with ≥1 CDI recurrence (diagnosed by PCR or toxin EIA, inclusive of the current episode). After standard-of-care antibiotics, subjects received SER-109 (4 capsules daily x 3 days). Efficacy endpoints, including the proportion of subjects with rCDI (toxin+ diarrhea requiring treatment), were evaluated through Week 8; safety and durability of response were assessed through Week 24.

Results: Of 351 subjects screened, 263 were enrolled (Cohort 1: N=29; Cohort 2: N=234; 68% female; mean age 64 years). Approximately 1/3 of subjects enrolled with their 1st recurrence. Comorbidities included cardiac disorders (31%), neoplasms (21%), Type 2 diabetes (11%), COPD (10%), chronic kidney disease (9%), and hepatobiliary disorders (9%). Overall, 141 subjects (54%) had treatment-emergent adverse events (TEAEs), the majority of which were mild to moderate and gastrointestinal. There were 8 deaths (3%) and 33 subjects (13%) with serious TEAEs; none were treatment-related (Table 1).

Overall, 240 subjects (91.3%) had clinical response (ie, no CDI recurrence) at 8 weeks. Of these, 227 subjects (94.6%) maintained a durable response through Week 24. Among the 240 responders at Week 8, 72 subjects (30%) had a history of 1st recurrence; the remainder had a history of ≥2 CDI recurrences. The proportion of Week 8 responders who maintained durable response through Week 24 was similar between those with a history of 1st recurrence vs ≥2 recurrences (94.4% vs 94.6%, respectively; Figure 1).

Conclusions: SER-109, a potential first-in-class oral investigational microbiome therapeutic, was well-tolerated in this population of patients with multiple comorbidities. High and durable clinical response rates were observed through Week 24, regardless of the number of prior CDI recurrences, supporting the potential benefit of microbiome repair following antibiotics to treat rCDI.

Background: Antibiotics used to treat CDI perturb the gut microbiome, increasing susceptibility to recurrent CDI (rCDI). Fecal microbiota transplants and donor-derived treatments promote the establishment of a gut environment resistant to CDI, but the composition and quality attributes of these are inherently variable; they are difficult to scale up, and have resulted in transfer of harmful pathogens. In contrast, VE303 is a rationally defined consortium that is manufactured from clonal cell banks, obviating the need for donors and overcoming the limitations of donor-derived treatments. In the Phase 2 CONSORTIUM Study, the VE303 high dose arm had an acceptable safety profile and significantly reduced the risk of rCDI compared with placebo. The high dose also induced superior VE303 strain colonization at 14 days, achieved long-term engraftment, and promoted early restoration of the microbiota and beneficial metabolites.

Methods: The CONSORTIUM Study was a randomized, double-blind, placebo-controlled, dose-finding study in individuals at high risk of rCDI. After completing a course of antibiotics for a lab-confirmed CDI episode, subjects were randomized 1:1:1 to low-dose VE303 (1.6 x 10⁸ CFU), high-dose VE303 (8 x 10⁸ CFU), or placebo orally once daily for 14 days. Subjects were followed for 24 weeks to monitor safety, rCDI episodes, and gut microbiota composition. Fecal samples were collected during dosing and at weeks 4 and 8; metagenomic sequencing and metabolomics analysis was performed to identify associations between VE303 strain engraftment, resident microbes, stool metabolites, and clinical safety and efficacy.

Results: Among 378 analyzed species, VE303-08, -02 and -01 were the only bacterial species found to be significantly associated with nonrecurrence (panel A, p-adjust<0.25, linear mixed effects [LME] model). Table 1 shows taxa associated with clinical response and VE303 at species and higher taxonomic levels, with greater Clostridia in nonrecurrent subjects and Proteobacteria in recurrent subjects. VE303 was negatively correlated (p-adjust<0.05, LME) with recurrence-associated taxa Veillonella and Kluyvera, and positively correlated with response-associated taxa. Also, multiple VE303 strains were positively correlated with gut metabolites (panel B, p-adjust<0.05, LME) including the short-chain fatty acids valerate, acetate and butyrate, and the 2° bile acid UDCA, which confers colonization resistance against C. difficile and inhibits spore germination.

Conclusion: In subjects at high risk of rCDI, VE303 species were positively correlated with beneficial Clostridia, short-chain fatty acids, and UDCA, and negatively correlated with recurrence-associated Proteobacteria. VE303 may protect against rCDI through a combination of direct exposure to the consortium strains and VE303-mediated recovery of the endogenous microbiota and metabolites.

<b>(A)</b> Effect-size bar plot for all species associated with recurrence or non-recurrence (LME, p-adjust < 0.25). Bars are colored according to the taxonomic class of each organism. Vertical legend indicates correlation with VE303. <b>(B)</b> Heatmap depicting the significant associations between stool SCFA and BA, and VE303 strain relative abundance (LME, p-adjust < 0.05).

(A) Effect-size bar plot for all species associated with recurrence or non-recurrence (LME, p-adjust < 0.25). Bars are colored according to the taxonomic class of each organism. Vertical legend indicates correlation with VE303. (B) Heatmap depicting the significant associations between stool SCFA and BA, and VE303 strain relative abundance (LME, p-adjust < 0.05).

Taxa associated with treatment response in VE303-dosed groups during the first 2 weeks post-antibiotics (unadjusted p < 0.05, LME). Positive or negative correlations with VE303 (adjusted p < 0.05, LME) are highlighted in blue and red respectively.

Taxa associated with treatment response in VE303-dosed groups during the first 2 weeks post-antibiotics (unadjusted p < 0.05, LME). Positive or negative correlations with VE303 (adjusted p < 0.05, LME) are highlighted in blue and red respectively.

Background: The prevalence of primary and recurrent Clostridioides difficile infection (CDI) has increased since 2000, but contributing factors remain unclear. Biofilm formation decreases C. difficile antibiotic susceptibility and can contribute to increased fitness in the gut. The aim of this study was to determine if patients with CDI are more likely to have colonic mucosal colonization with C. difficile.
Methods: We identified 52 patients with primary episode of CDI and a matched group of 84 patients without CDI for whom archived colonic biopsies (formalin fixed, paraffin embedded) were collected within one year prior (range 7-363; median 134). We assessed presence or absence of C. difficile biofilm in biopsies by looking for overlap of universal bacterial 16S rRNA (EUB338 and C. difficile-specific 16S rRNA fluorescent in-situ hybridization (FISH) probe) and C. difficile biofilm polysaccharide II (PSII; immunohistochemistry). To identify factors contributing to CDI we used univariate and multivariate generalized logistic regression analysis (GLM; R stats V 3.6.2 package). To quantify C. difficile biofilm in colonic mucosa, signals from DAPI and MUC2 channels identified background host tissue and a custom script contoured the host tissue. The same mask was applied to images in PSII, EUB338, and C. difficile 16S FISH channels. The overlap of pixels detected in EUB338, C. difficile 16S FISH and PSII was considered positive for C. difficile biofilm. Since biofilms may position around C. difficile cells in unpredictable locations, the PSII detection range extended to 10µm around the cell to allow identification of biofilm matrix. Biofilm detection algorithm was tested on 10 biofilm-positive or 10 negative samples examined and sorted manually. There was statistically significant higher pixel count in biofilm-positive samples. A pixel count cutoff of 100 separated biofilm-positive samples from biofilm-negative samples.
Results: There were no significant differences in age, sex, BMI, or inflammatory bowel disease among CDI and non-CDI groups (Table 1). C. difficile biofilm was identified in 30/51 (58.8%) patients prior to primary episode of CDI and 5/84 (5.6%) patients without CDI. Univariate analysis showed patients with CDI were more likely to have antibiotic use, hospitalization, corticosteroid use (trend), and colonic mucosal colonization as evidenced by C. difficile biofilm (Table 2). A multivariate analysis showed patients with CDI were more likely to have hospitalization, corticosteroid use (trend) and C. difficile biofilm (Table 2).
Conclusion: Our findings suggest asymptomatic individuals can harbor C. difficile in the colonic mucosa and such colonization may increase risk of future CDI. Larger prospective studies are needed to confirm the role of C. difficile biofilm formation in primary and recurrent CDI.

Robin Patel

Appearances